[ViewVC] Diff of: cvs/CBOR-XS/XS.xs

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.20 by root, Wed Nov 20 11:06:42 2013 UTC vs.
Revision 1.50 by root, Thu Feb 25 02:29:22 2016 UTC

…		…
7	#include <stdlib.h>	7	#include <stdlib.h>
8	#include <stdio.h>	8	#include <stdio.h>
9	#include <limits.h>	9	#include <limits.h>
10	#include <float.h>	10	#include <float.h>
11		11
		12	#define ECB_NO_THREADS 1
12	#include "ecb.h"	13	#include "ecb.h"
13		14
14	// compatibility with perl <5.18	15	// compatibility with perl <5.18
15	#ifndef HvNAMELEN_get	16	#ifndef HvNAMELEN_get
16	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))	17	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
…		…
19	# define HvNAMELEN(hv) HvNAMELEN_get (hv)	20	# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20	#endif	21	#endif
21	#ifndef HvNAMEUTF8	22	#ifndef HvNAMEUTF8
22	# define HvNAMEUTF8(hv) 0	23	# define HvNAMEUTF8(hv) 0
23	#endif	24	#endif
		25	#ifndef SvREFCNT_dec_NN
		26	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
		27	#endif
		28
		29	// known major and minor types
		30	enum cbor_type
		31	{
		32	MAJOR_SHIFT = 5,
		33	MINOR_MASK = 0x1f,
		34
		35	MAJOR_POS_INT = 0 << MAJOR_SHIFT,
		36	MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
		37	MAJOR_BYTES = 2 << MAJOR_SHIFT,
		38	MAJOR_TEXT = 3 << MAJOR_SHIFT,
		39	MAJOR_ARRAY = 4 << MAJOR_SHIFT,
		40	MAJOR_MAP = 5 << MAJOR_SHIFT,
		41	MAJOR_TAG = 6 << MAJOR_SHIFT,
		42	MAJOR_MISC = 7 << MAJOR_SHIFT,
		43
		44	// INT/STRING/ARRAY/MAP subtypes
		45	LENGTH_EXT1 = 24,
		46	LENGTH_EXT2 = 25,
		47	LENGTH_EXT4 = 26,
		48	LENGTH_EXT8 = 27,
		49
		50	// SIMPLE types (effectively MISC subtypes)
		51	SIMPLE_FALSE = 20,
		52	SIMPLE_TRUE = 21,
		53	SIMPLE_NULL = 22,
		54	SIMPLE_UNDEF = 23,
		55
		56	// MISC subtype (unused)
		57	MISC_EXT1 = 24,
		58	MISC_FLOAT16 = 25,
		59	MISC_FLOAT32 = 26,
		60	MISC_FLOAT64 = 27,
		61
		62	// BYTES/TEXT/ARRAY/MAP
		63	MINOR_INDEF = 31,
		64	};
24		65
25	// known tags	66	// known tags
26	enum cbor_tag	67	enum cbor_tag
27	{	68	{
28	// inofficial extensions (pending iana registration)	69	// extensions
		70	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29	CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object	71	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30	CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object	72	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31	CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing	73	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32	CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33	CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref	75	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34	CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	76	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36		77
37	// rfc7049	78	// rfc7049
38	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	79	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	80	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40	CBOR_TAG_POS_BIGNUM = 2, // byte string	81	CBOR_TAG_POS_BIGNUM = 2, // byte string
41	CBOR_TAG_NEG_BIGNUM = 3, // byte string	82	CBOR_TAG_NEG_BIGNUM = 3, // byte string
42	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	83	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43	CBOR_TAG_BIGFLOAT = 5, // array	84	CBOR_TAG_BIGFLOAT = 5, // array
44		85
45	CBOR_TAG_CONV_B64U = 21, // base64url, any	86	CBOR_TAG_CONV_B64U = 21, // base64url, any
46	CBOR_TAG_CONV_B64 = 22, // base64, any	87	CBOR_TAG_CONV_B64 = 22, // base64, any
47	CBOR_TAG_CONV_HEX = 23, // base16, any	88	CBOR_TAG_CONV_HEX = 23, // base16, any
48	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	89	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49		90
50	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	91	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	92	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	93	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	94	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	95	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55		96
56	CBOR_TAG_MAGIC = 55799 // self-describe cbor	97	CBOR_TAG_MAGIC = 55799, // self-describe cbor
57	};	98	};
58		99
59	#define F_SHRINK 0x00000001UL	100	#define F_SHRINK 0x00000001UL
60	#define F_ALLOW_UNKNOWN 0x00000002UL	101	#define F_ALLOW_UNKNOWN 0x00000002UL
61	#define F_ALLOW_SHARING 0x00000004UL //TODO	102	#define F_ALLOW_SHARING 0x00000004UL
		103	#define F_ALLOW_CYCLES 0x00000008UL
62	#define F_DEDUP_STRINGS 0x00000008UL //TODO	104	#define F_PACK_STRINGS 0x00000010UL
63	#define F_DEDUP_KEYS 0x00000010UL //TODO	105	#define F_VALIDATE_UTF8 0x00000020UL
64		106
65	#define INIT_SIZE 32 // initial scalar size to be allocated	107	#define INIT_SIZE 32 // initial scalar size to be allocated
66		108
67	#define SB do {	109	#define SB do {
68	#define SE } while (0)	110	#define SE } while (0)
…		…
80	# define CBOR_SLOW 0	122	# define CBOR_SLOW 0
81	# define CBOR_STASH cbor_stash	123	# define CBOR_STASH cbor_stash
82	#endif	124	#endif
83		125
84	static HV cbor_stash, types_boolean_stash, types_error_stash, cbor_tagged_stash; // CBOR::XS::	126	static HV cbor_stash, types_boolean_stash, types_error_stash, cbor_tagged_stash; // CBOR::XS::
85	static SV types_true, types_false, types_error, sv_cbor;	127	static SV types_true, types_false, types_error, sv_cbor, *default_filter;
86		128
87	typedef struct {	129	typedef struct {
88	U32 flags;	130	U32 flags;
89	U32 max_depth;	131	U32 max_depth;
90	STRLEN max_size;	132	STRLEN max_size;
		133	SV *filter;
		134
		135	// for the incremental parser
		136	STRLEN incr_pos; // the current offset into the text
		137	STRLEN incr_need; // minimum bytes needed to decode
		138	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
91	} CBOR;	139	} CBOR;
92		140
93	ecb_inline void	141	ecb_inline void
94	cbor_init (CBOR *cbor)	142	cbor_init (CBOR *cbor)
95	{	143	{
96	Zero (cbor, 1, CBOR);	144	Zero (cbor, 1, CBOR);
97	cbor->max_depth = 512;	145	cbor->max_depth = 512;
		146	}
		147
		148	ecb_inline void
		149	cbor_free (CBOR *cbor)
		150	{
		151	SvREFCNT_dec (cbor->filter);
		152	SvREFCNT_dec (cbor->incr_count);
98	}	153	}
99		154
100	/////////////////////////////////////////////////////////////////////////////	155	/////////////////////////////////////////////////////////////////////////////
101	// utility functions	156	// utility functions
102		157
…		…
122	SvPV_shrink_to_cur (sv);	177	SvPV_shrink_to_cur (sv);
123	#elif defined (SvPV_renew)	178	#elif defined (SvPV_renew)
124	SvPV_renew (sv, SvCUR (sv) + 1);	179	SvPV_renew (sv, SvCUR (sv) + 1);
125	#endif	180	#endif
126	}	181	}
		182	}
		183
		184	// minimum length of a string to be registered for stringref
		185	ecb_inline int
		186	minimum_string_length (UV idx)
		187	{
		188	return idx > 23
		189	? idx > 0xffU
		190	? idx > 0xffffU
		191	? idx > 0xffffffffU
		192	? 11
		193	: 7
		194	: 5
		195	: 4
		196	: 3;
127	}	197	}
128		198
129	/////////////////////////////////////////////////////////////////////////////	199	/////////////////////////////////////////////////////////////////////////////
130	// encoder	200	// encoder
131		201
…		…
165	static void	235	static void
166	encode_uint (enc_t *enc, int major, UV len)	236	encode_uint (enc_t *enc, int major, UV len)
167	{	237	{
168	need (enc, 9);	238	need (enc, 9);
169		239
170	if (len < 24)	240	if (ecb_expect_true (len < LENGTH_EXT1))
171	*enc->cur++ = major \| len;	241	*enc->cur++ = major \| len;
172	else if (len <= 0xff)	242	else if (ecb_expect_true (len <= 0xffU))
173	{	243	{
174	*enc->cur++ = major \| 24;	244	*enc->cur++ = major \| LENGTH_EXT1;
175	*enc->cur++ = len;	245	*enc->cur++ = len;
176	}	246	}
177	else if (len <= 0xffff)	247	else if (len <= 0xffffU)
178	{	248	{
179	*enc->cur++ = major \| 25;	249	*enc->cur++ = major \| LENGTH_EXT2;
180	*enc->cur++ = len >> 8;	250	*enc->cur++ = len >> 8;
181	*enc->cur++ = len;	251	*enc->cur++ = len;
182	}	252	}
183	else if (len <= 0xffffffff)	253	else if (len <= 0xffffffffU)
184	{	254	{
185	*enc->cur++ = major \| 26;	255	*enc->cur++ = major \| LENGTH_EXT4;
186	*enc->cur++ = len >> 24;	256	*enc->cur++ = len >> 24;
187	*enc->cur++ = len >> 16;	257	*enc->cur++ = len >> 16;
188	*enc->cur++ = len >> 8;	258	*enc->cur++ = len >> 8;
189	*enc->cur++ = len;	259	*enc->cur++ = len;
190	}	260	}
191	else	261	else
192	{	262	{
193	*enc->cur++ = major \| 27;	263	*enc->cur++ = major \| LENGTH_EXT8;
194	*enc->cur++ = len >> 56;	264	*enc->cur++ = len >> 56;
195	*enc->cur++ = len >> 48;	265	*enc->cur++ = len >> 48;
196	*enc->cur++ = len >> 40;	266	*enc->cur++ = len >> 40;
197	*enc->cur++ = len >> 32;	267	*enc->cur++ = len >> 32;
198	*enc->cur++ = len >> 24;	268	*enc->cur++ = len >> 24;
…		…
200	*enc->cur++ = len >> 8;	270	*enc->cur++ = len >> 8;
201	*enc->cur++ = len;	271	*enc->cur++ = len;
202	}	272	}
203	}	273	}
204		274
205	static void	275	ecb_inline void
		276	encode_tag (enc_t *enc, UV tag)
		277	{
		278	encode_uint (enc, MAJOR_TAG, tag);
		279	}
		280
		281	ecb_inline void
206	encode_str (enc_t enc, int utf8, char str, STRLEN len)	282	encode_str (enc_t enc, int utf8, char str, STRLEN len)
207	{	283	{
208	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	284	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
209	need (enc, len);	285	need (enc, len);
210	memcpy (enc->cur, str, len);	286	memcpy (enc->cur, str, len);
211	enc->cur += len;	287	enc->cur += len;
212	}	288	}
213		289
214	ecb_inline void	290	static void
215	encode_tag (enc_t *enc, UV tag)	291	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
216	{	292	{
217	encode_uint (enc, 0xc0, tag);	293	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
		294	{
		295	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
		296
		297	if (SvOK (*svp))
		298	{
		299	// already registered, use stringref
		300	encode_tag (enc, CBOR_TAG_STRINGREF);
		301	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		302	return;
		303	}
		304	else if (len >= minimum_string_length (enc->stringref_idx))
		305	{
		306	// register only
		307	sv_setuv (*svp, enc->stringref_idx);
		308	++enc->stringref_idx;
		309	}
		310	}
		311
		312	encode_str (enc, utf8, str, len);
218	}	313	}
219		314
220	static void encode_sv (enc_t enc, SV sv);	315	static void encode_sv (enc_t enc, SV sv);
221		316
222	static void	317	static void
…		…
227	if (enc->depth >= enc->cbor.max_depth)	322	if (enc->depth >= enc->cbor.max_depth)
228	croak (ERR_NESTING_EXCEEDED);	323	croak (ERR_NESTING_EXCEEDED);
229		324
230	++enc->depth;	325	++enc->depth;
231		326
232	encode_uint (enc, 0x80, len + 1);	327	encode_uint (enc, MAJOR_ARRAY, len + 1);
233		328
		329	if (SvMAGICAL (av))
234	for (i = 0; i <= len; ++i)	330	for (i = 0; i <= len; ++i)
235	{	331	{
236	SV **svp = av_fetch (av, i, 0);	332	SV **svp = av_fetch (av, i, 0);
237	encode_sv (enc, svp ? *svp : &PL_sv_undef);	333	encode_sv (enc, svp ? *svp : &PL_sv_undef);
238	}	334	}
		335	else
		336	for (i = 0; i <= len; ++i)
		337	{
		338	SV *sv = AvARRAY (av)[i];
		339	encode_sv (enc, sv ? sv : &PL_sv_undef);
		340	}
239		341
240	--enc->depth;	342	--enc->depth;
241	}
242
243	ecb_inline void
244	encode_he (enc_t enc, HE he)
245	{
246	if (HeKLEN (he) == HEf_SVKEY)
247	encode_sv (enc, HeSVKEY (he));
248	else
249	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
250	}	343	}
251		344
252	static void	345	static void
253	encode_hv (enc_t enc, HV hv)	346	encode_hv (enc_t enc, HV hv)
254	{	347	{
…		…
261		354
262	int pairs = hv_iterinit (hv);	355	int pairs = hv_iterinit (hv);
263	int mg = SvMAGICAL (hv);	356	int mg = SvMAGICAL (hv);
264		357
265	if (mg)	358	if (mg)
266	encode_ch (enc, 0xa0 \| 31);	359	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
267	else	360	else
268	encode_uint (enc, 0xa0, pairs);	361	encode_uint (enc, MAJOR_MAP, pairs);
269		362
270	while ((he = hv_iternext (hv)))	363	while ((he = hv_iternext (hv)))
271	{	364	{
272	if (ecb_expect_false (enc->cbor.flags & (F_DEDUP_STRINGS \| F_DEDUP_KEYS)))
273	{
274	SV **svp;
275
276	if (HeKLEN (he) == HEf_SVKEY)	365	if (HeKLEN (he) == HEf_SVKEY)
277	svp = hv_fetch_ent (enc->stringref[!! SvUTF8 (HeSVKEY (he))], HeSVKEY (he) , 1, 0);//TODO return HE :/	366	encode_sv (enc, HeSVKEY (he));
278	else
279	svp = hv_fetch (enc->stringref[!! HeKUTF8 (he) ], HeKEY (he), HeKLEN (he), 1);
280
281	if (SvOK (*svp))
282	{
283	encode_tag (enc, CBOR_TAG_STRINGREF);
284	encode_uint (enc, 0x00, SvUV (*svp));
285	}
286	else
287	{
288	sv_setuv (*svp, enc->stringref_idx);
289	++enc->stringref_idx;
290	encode_he (enc, he);
291	}
292	}
293	else	367	else
294	encode_he (enc, he);	368	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
295		369
296	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	370	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
297	}	371	}
298		372
299	if (mg)	373	if (mg)
300	encode_ch (enc, 0xe0 \| 31);	374	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
301		375
302	--enc->depth;	376	--enc->depth;
303	}	377	}
304		378
305	// encode objects, arrays and special \0=false and \1=true values.	379	// encode objects, arrays and special \0=false and \1=true values.
306	static void	380	static void
307	encode_rv (enc_t enc, SV sv)	381	encode_rv (enc_t enc, SV sv)
308	{	382	{
309	SvGETMAGIC (sv);	383	SvGETMAGIC (sv);
310
311	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
312	&& ecb_expect_false (SvREFCNT (sv) > 1))
313	{
314	if (!enc->shareable)
315	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
316
317	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
318
319	if (SvOK (*svp))
320	{
321	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
322	encode_uint (enc, 0x00, SvUV (*svp));
323	return;
324	}
325	else
326	{
327	sv_setuv (*svp, enc->shareable_idx);
328	++enc->shareable_idx;
329	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
330	}
331	}
332		384
333	svtype svt = SvTYPE (sv);	385	svtype svt = SvTYPE (sv);
334		386
335	if (ecb_expect_false (SvOBJECT (sv)))	387	if (ecb_expect_false (SvOBJECT (sv)))
336	{	388	{
…		…
343	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	395	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
344	? cbor_tagged_stash	396	? cbor_tagged_stash
345	: gv_stashpv ("CBOR::XS::Tagged" , 1);	397	: gv_stashpv ("CBOR::XS::Tagged" , 1);
346		398
347	HV *stash = SvSTASH (sv);	399	HV *stash = SvSTASH (sv);
348	GV *method;
349		400
350	if (stash == boolean_stash)	401	if (stash == boolean_stash)
351	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	402	{
		403	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		404	return;
		405	}
352	else if (stash == error_stash)	406	else if (stash == error_stash)
353	encode_ch (enc, 0xe0 \| 23);	407	{
		408	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		409	return;
		410	}
354	else if (stash == tagged_stash)	411	else if (stash == tagged_stash)
355	{	412	{
356	if (svt != SVt_PVAV)	413	if (svt != SVt_PVAV)
357	croak ("encountered CBOR::XS::Tagged object that isn't an array");	414	croak ("encountered CBOR::XS::Tagged object that isn't an array");
358		415
359	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	416	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
360	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	417	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		418
		419	return;
		420	}
		421	}
		422
		423	if (ecb_expect_false (SvREFCNT (sv) > 1)
		424	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		425	{
		426	if (!enc->shareable)
		427	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		428
		429	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		430
		431	if (SvOK (*svp))
361	}	432	{
		433	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		434	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		435	return;
		436	}
		437	else
		438	{
		439	sv_setuv (*svp, enc->shareable_idx);
		440	++enc->shareable_idx;
		441	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		442	}
		443	}
		444
		445	if (ecb_expect_false (SvOBJECT (sv)))
		446	{
		447	HV *stash = SvSTASH (sv);
		448	GV *method;
		449
362	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	450	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
363	{	451	{
364	dSP;	452	dSP;
365		453
366	ENTER; SAVETMPS; PUSHMARK (SP);	454	ENTER; SAVETMPS;
		455	PUSHMARK (SP);
367	// we re-bless the reference to get overload and other niceties right	456	// we re-bless the reference to get overload and other niceties right
368	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	457	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
369		458
370	PUTBACK;	459	PUTBACK;
371	// G_SCALAR ensures that return value is 1	460	// G_SCALAR ensures that return value is 1
…		…
384	}	473	}
385	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	474	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
386	{	475	{
387	dSP;	476	dSP;
388		477
389	ENTER; SAVETMPS; PUSHMARK (SP);	478	ENTER; SAVETMPS;
		479	SAVESTACK_POS ();
		480	PUSHMARK (SP);
390	EXTEND (SP, 2);	481	EXTEND (SP, 2);
391	// we re-bless the reference to get overload and other niceties right	482	// we re-bless the reference to get overload and other niceties right
392	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	483	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
393	PUSHs (sv_cbor);	484	PUSHs (sv_cbor);
394		485
…		…
399	// catch this surprisingly common error	490	// catch this surprisingly common error
400	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	491	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
401	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	492	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
402		493
403	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	494	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
404	encode_uint (enc, 0x80, count + 1);	495	encode_uint (enc, MAJOR_ARRAY, count + 1);
405	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	496	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
406		497
407	while (count)	498	while (count)
408	encode_sv (enc, SP[1 - count--]);	499	encode_sv (enc, SP[1 - count--]);
409		500
410	PUTBACK;	501	PUTBACK;
…		…
431	{	522	{
432	double nv = SvNVX (sv);	523	double nv = SvNVX (sv);
433		524
434	need (enc, 9);	525	need (enc, 9);
435		526
436	if (ecb_expect_false (nv == (U32)nv))	527	if (ecb_expect_false (nv == (NV)(U32)nv))
437	encode_uint (enc, 0x00, (U32)nv);	528	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
438	//TODO: maybe I32?	529	//TODO: maybe I32?
439	else if (ecb_expect_false (nv == (float)nv))	530	else if (ecb_expect_false (nv == (float)nv))
440	{	531	{
441	uint32_t fp = ecb_float_to_binary32 (nv);	532	uint32_t fp = ecb_float_to_binary32 (nv);
442		533
443	*enc->cur++ = 0xe0 \| 26;	534	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
444		535
445	if (!ecb_big_endian ())	536	if (!ecb_big_endian ())
446	fp = ecb_bswap32 (fp);	537	fp = ecb_bswap32 (fp);
447		538
448	memcpy (enc->cur, &fp, 4);	539	memcpy (enc->cur, &fp, 4);
…		…
450	}	541	}
451	else	542	else
452	{	543	{
453	uint64_t fp = ecb_double_to_binary64 (nv);	544	uint64_t fp = ecb_double_to_binary64 (nv);
454		545
455	*enc->cur++ = 0xe0 \| 27;	546	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
456		547
457	if (!ecb_big_endian ())	548	if (!ecb_big_endian ())
458	fp = ecb_bswap64 (fp);	549	fp = ecb_bswap64 (fp);
459		550
460	memcpy (enc->cur, &fp, 8);	551	memcpy (enc->cur, &fp, 8);
…		…
469		560
470	if (SvPOKp (sv))	561	if (SvPOKp (sv))
471	{	562	{
472	STRLEN len;	563	STRLEN len;
473	char *str = SvPV (sv, len);	564	char *str = SvPV (sv, len);
474	encode_str (enc, SvUTF8 (sv), str, len);	565	encode_strref (enc, SvUTF8 (sv), str, len);
475	}	566	}
476	else if (SvNOKp (sv))	567	else if (SvNOKp (sv))
477	encode_nv (enc, sv);	568	encode_nv (enc, sv);
478	else if (SvIOKp (sv))	569	else if (SvIOKp (sv))
479	{	570	{
480	if (SvIsUV (sv))	571	if (SvIsUV (sv))
481	encode_uint (enc, 0x00, SvUVX (sv));	572	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
482	else if (SvIVX (sv) >= 0)	573	else if (SvIVX (sv) >= 0)
483	encode_uint (enc, 0x00, SvIVX (sv));	574	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
484	else	575	else
485	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	576	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
486	}	577	}
487	else if (SvROK (sv))	578	else if (SvROK (sv))
488	encode_rv (enc, SvRV (sv));	579	encode_rv (enc, SvRV (sv));
489	else if (!SvOK (sv))	580	else if (!SvOK (sv))
490	encode_ch (enc, 0xe0 \| 22);	581	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
491	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	582	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
492	encode_ch (enc, 0xe0 \| 23);	583	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
493	else	584	else
494	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	585	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
495	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	586	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
496	}	587	}
497		588
498	static SV *	589	static SV *
499	encode_cbor (SV scalar, CBOR cbor)	590	encode_cbor (SV scalar, CBOR cbor)
500	{	591	{
501	enc_t enc = { };	592	enc_t enc = { 0 };
502		593
503	enc.cbor = *cbor;	594	enc.cbor = *cbor;
504	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	595	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
505	enc.cur = SvPVX (enc.sv);	596	enc.cur = SvPVX (enc.sv);
506	enc.end = SvEND (enc.sv);	597	enc.end = SvEND (enc.sv);
507		598
508	SvPOK_only (enc.sv);	599	SvPOK_only (enc.sv);
509		600
510	if (cbor->flags & (F_DEDUP_STRINGS \| F_DEDUP_KEYS))	601	if (cbor->flags & F_PACK_STRINGS)
511	{	602	{
512	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	603	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
513	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	604	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
514	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	605	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
515	}	606	}
…		…
537	CBOR cbor;	628	CBOR cbor;
538	U32 depth; // recursion depth	629	U32 depth; // recursion depth
539	U32 maxdepth; // recursion depth limit	630	U32 maxdepth; // recursion depth limit
540	AV *shareable;	631	AV *shareable;
541	AV *stringref;	632	AV *stringref;
		633	SV *decode_tagged;
542	} dec_t;	634	} dec_t;
543		635
544	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	636	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
545		637
546	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	638	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
…		…
549	#define DEC_DEC_DEPTH --dec->depth	641	#define DEC_DEC_DEPTH --dec->depth
550		642
551	static UV	643	static UV
552	decode_uint (dec_t *dec)	644	decode_uint (dec_t *dec)
553	{	645	{
554	switch (*dec->cur & 31)	646	U8 m = *dec->cur & MINOR_MASK;
555	{	647	++dec->cur;
556	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
557	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
558	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
559	return *dec->cur++ & 31;
560		648
561	case 24:	649	if (ecb_expect_true (m < LENGTH_EXT1))
		650	return m;
		651	else if (ecb_expect_true (m == LENGTH_EXT1))
		652	{
562	WANT (2);	653	WANT (1);
563	dec->cur += 2;	654	dec->cur += 1;
564	return dec->cur[-1];	655	return dec->cur[-1];
565		656	}
566	case 25:	657	else if (ecb_expect_true (m == LENGTH_EXT2))
		658	{
567	WANT (3);	659	WANT (2);
568	dec->cur += 3;	660	dec->cur += 2;
569	return (((UV)dec->cur[-2]) << 8)	661	return (((UV)dec->cur[-2]) << 8)
570	\| ((UV)dec->cur[-1]);	662	\| ((UV)dec->cur[-1]);
571		663	}
572	case 26:	664	else if (ecb_expect_true (m == LENGTH_EXT4))
		665	{
573	WANT (5);	666	WANT (4);
574	dec->cur += 5;	667	dec->cur += 4;
575	return (((UV)dec->cur[-4]) << 24)	668	return (((UV)dec->cur[-4]) << 24)
576	\| (((UV)dec->cur[-3]) << 16)	669	\| (((UV)dec->cur[-3]) << 16)
577	\| (((UV)dec->cur[-2]) << 8)	670	\| (((UV)dec->cur[-2]) << 8)
578	\| ((UV)dec->cur[-1]);	671	\| ((UV)dec->cur[-1]);
579		672	}
580	case 27:	673	else if (ecb_expect_true (m == LENGTH_EXT8))
		674	{
581	WANT (9);	675	WANT (8);
582	dec->cur += 9;	676	dec->cur += 8;
		677
		678	return
		679	#if UVSIZE < 8
		680	0
		681	#else
583	return (((UV)dec->cur[-8]) << 56)	682	(((UV)dec->cur[-8]) << 56)
584	\| (((UV)dec->cur[-7]) << 48)	683	\| (((UV)dec->cur[-7]) << 48)
585	\| (((UV)dec->cur[-6]) << 40)	684	\| (((UV)dec->cur[-6]) << 40)
586	\| (((UV)dec->cur[-5]) << 32)	685	\| (((UV)dec->cur[-5]) << 32)
		686	#endif
587	\| (((UV)dec->cur[-4]) << 24)	687	\| (((UV)dec->cur[-4]) << 24)
588	\| (((UV)dec->cur[-3]) << 16)	688	\| (((UV)dec->cur[-3]) << 16)
589	\| (((UV)dec->cur[-2]) << 8)	689	\| (((UV)dec->cur[-2]) << 8)
590	\| ((UV)dec->cur[-1]);	690	\| ((UV)dec->cur[-1]);
591		691	}
592	default:	692	else
593	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	693	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
594	}
595		694
596	fail:	695	fail:
597	return 0;	696	return 0;
598	}	697	}
599		698
…		…
604	{	703	{
605	AV *av = newAV ();	704	AV *av = newAV ();
606		705
607	DEC_INC_DEPTH;	706	DEC_INC_DEPTH;
608		707
609	if ((*dec->cur & 31) == 31)	708	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
610	{	709	{
611	++dec->cur;	710	++dec->cur;
612		711
613	for (;;)	712	for (;;)
614	{	713	{
615	WANT (1);	714	WANT (1);
616		715
617	if (*dec->cur == (0xe0 \| 31))	716	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
618	{	717	{
619	++dec->cur;	718	++dec->cur;
620	break;	719	break;
621	}	720	}
622		721
…		…
625	}	724	}
626	else	725	else
627	{	726	{
628	int i, len = decode_uint (dec);	727	int i, len = decode_uint (dec);
629		728
		729	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
630	av_fill (av, len - 1);	730	av_fill (av, len - 1);
631		731
632	for (i = 0; i < len; ++i)	732	for (i = 0; i < len; ++i)
633	AvARRAY (av)[i] = decode_sv (dec);	733	AvARRAY (av)[i] = decode_sv (dec);
634	}	734	}
…		…
644		744
645	static void	745	static void
646	decode_he (dec_t dec, HV hv)	746	decode_he (dec_t dec, HV hv)
647	{	747	{
648	// for speed reasons, we specialcase single-string	748	// for speed reasons, we specialcase single-string
649	// byte or utf-8 strings as keys.	749	// byte or utf-8 strings as keys, but only when !stringref
650		750
651	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	751	if (ecb_expect_true (!dec->stringref))
		752	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
652	{	753	{
653	I32 len = decode_uint (dec);	754	I32 len = decode_uint (dec);
654	char key = (char )dec->cur;	755	char key = (char )dec->cur;
655		756
		757	WANT (len);
656	dec->cur += len;	758	dec->cur += len;
657		759
658	if (ecb_expect_false (dec->stringref))
659	av_push (dec->stringref, newSVpvn (key, len));
660
661	hv_store (hv, key, len, decode_sv (dec), 0);	760	hv_store (hv, key, len, decode_sv (dec), 0);
		761
		762	return;
662	}	763	}
663	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	764	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
664	{	765	{
665	I32 len = decode_uint (dec);	766	I32 len = decode_uint (dec);
666	char key = (char )dec->cur;	767	char key = (char )dec->cur;
667		768
		769	WANT (len);
668	dec->cur += len;	770	dec->cur += len;
669		771
670	if (ecb_expect_false (dec->stringref))	772	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
671	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	773	if (!is_utf8_string (key, len))
		774	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
672		775
673	hv_store (hv, key, -len, decode_sv (dec), 0);	776	hv_store (hv, key, -len, decode_sv (dec), 0);
		777
		778	return;
674	}	779	}
675	else	780
676	{
677	SV *k = decode_sv (dec);	781	SV *k = decode_sv (dec);
678	SV *v = decode_sv (dec);	782	SV *v = decode_sv (dec);
679		783
680	hv_store_ent (hv, k, v, 0);	784	hv_store_ent (hv, k, v, 0);
681	SvREFCNT_dec (k);	785	SvREFCNT_dec (k);
682	}	786
		787	fail:
		788	;
683	}	789	}
684		790
685	static SV *	791	static SV *
686	decode_hv (dec_t *dec)	792	decode_hv (dec_t *dec)
687	{	793	{
688	HV *hv = newHV ();	794	HV *hv = newHV ();
689		795
690	DEC_INC_DEPTH;	796	DEC_INC_DEPTH;
691		797
692	if ((*dec->cur & 31) == 31)	798	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
693	{	799	{
694	++dec->cur;	800	++dec->cur;
695		801
696	for (;;)	802	for (;;)
697	{	803	{
698	WANT (1);	804	WANT (1);
699		805
700	if (*dec->cur == (0xe0 \| 31))	806	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
701	{	807	{
702	++dec->cur;	808	++dec->cur;
703	break;	809	break;
704	}	810	}
705		811
…		…
726	static SV *	832	static SV *
727	decode_str (dec_t *dec, int utf8)	833	decode_str (dec_t *dec, int utf8)
728	{	834	{
729	SV *sv = 0;	835	SV *sv = 0;
730		836
731	if ((*dec->cur & 31) == 31)	837	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
732	{	838	{
		839	// indefinite length strings
733	++dec->cur;	840	++dec->cur;
734		841
		842	U8 major = *dec->cur & MAJOR_MISC;
		843
735	sv = newSVpvn ("", 0);	844	sv = newSVpvn ("", 0);
736		845
737	// not very fast, and certainly not robust against illegal input
738	for (;;)	846	for (;;)
739	{	847	{
740	WANT (1);	848	WANT (1);
741		849
742	if (*dec->cur == (0xe0 \| 31))	850	if ((*dec->cur - major) > LENGTH_EXT8)
		851	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
743	{	852	{
744	++dec->cur;	853	++dec->cur;
745	break;	854	break;
746	}	855	}
		856	else
		857	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
747		858
748	sv_catsv (sv, decode_sv (dec));	859	STRLEN len = decode_uint (dec);
		860
		861	WANT (len);
		862	sv_catpvn (sv, dec->cur, len);
		863	dec->cur += len;
749	}	864	}
750	}	865	}
751	else	866	else
752	{	867	{
753	STRLEN len = decode_uint (dec);	868	STRLEN len = decode_uint (dec);
754		869
755	WANT (len);	870	WANT (len);
756	sv = newSVpvn (dec->cur, len);	871	sv = newSVpvn (dec->cur, len);
757	dec->cur += len;	872	dec->cur += len;
		873
		874	if (ecb_expect_false (dec->stringref)
		875	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
		876	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
758	}	877	}
759		878
760	if (utf8)	879	if (utf8)
		880	{
		881	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		882	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		883	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		884
761	SvUTF8_on (sv);	885	SvUTF8_on (sv);
762		886	}
763	if (ecb_expect_false (dec->stringref))
764	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
765		887
766	return sv;	888	return sv;
767		889
768	fail:	890	fail:
769	SvREFCNT_dec (sv);	891	SvREFCNT_dec (sv);
…		…
801	}	923	}
802	break;	924	break;
803		925
804	case CBOR_TAG_STRINGREF:	926	case CBOR_TAG_STRINGREF:
805	{	927	{
806	if ((*dec->cur >> 5) != 0)	928	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
807	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	929	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
808		930
809	UV idx = decode_uint (dec);	931	UV idx = decode_uint (dec);
810		932
811	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	933	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
818	case CBOR_TAG_VALUE_SHAREABLE:	940	case CBOR_TAG_VALUE_SHAREABLE:
819	{	941	{
820	if (ecb_expect_false (!dec->shareable))	942	if (ecb_expect_false (!dec->shareable))
821	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	943	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
822		944
		945	if (dec->cbor.flags & F_ALLOW_CYCLES)
		946	{
823	sv = newSV (0);	947	sv = newSV (0);
824	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	948	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
825		949
826	SV *osv = decode_sv (dec);	950	SV *osv = decode_sv (dec);
827	sv_setsv (sv, osv);	951	sv_setsv (sv, osv);
828	SvREFCNT_dec_NN (osv);	952	SvREFCNT_dec_NN (osv);
		953	}
		954	else
		955	{
		956	av_push (dec->shareable, &PL_sv_undef);
		957	int idx = AvFILLp (dec->shareable);
		958	sv = decode_sv (dec);
		959	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		960	}
829	}	961	}
830	break;	962	break;
831		963
832	case CBOR_TAG_VALUE_SHAREDREF:	964	case CBOR_TAG_VALUE_SHAREDREF:
833	{	965	{
834	if ((*dec->cur >> 5) != 0)	966	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
835	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	967	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
836		968
837	UV idx = decode_uint (dec);	969	UV idx = decode_uint (dec);
838		970
839	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	971	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
840	ERR ("corrupted CBOR data (sharedref index out of bounds)");	972	ERR ("corrupted CBOR data (sharedref index out of bounds)");
841		973
842	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	974	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		975
		976	if (sv == &PL_sv_undef)
		977	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
843	}	978	}
844	break;	979	break;
845		980
846	case CBOR_TAG_PERL_OBJECT:	981	case CBOR_TAG_PERL_OBJECT:
847	{	982	{
…		…
862	if (!method)	997	if (!method)
863	ERR ("cannot decode perl-object (package does not have a THAW method)");	998	ERR ("cannot decode perl-object (package does not have a THAW method)");
864		999
865	dSP;	1000	dSP;
866		1001
867	ENTER; SAVETMPS; PUSHMARK (SP);	1002	ENTER; SAVETMPS;
		1003	PUSHMARK (SP);
868	EXTEND (SP, len + 1);	1004	EXTEND (SP, len + 1);
869	// we re-bless the reference to get overload and other niceties right	1005	// we re-bless the reference to get overload and other niceties right
870	PUSHs (*av_fetch (av, 0, 1));	1006	PUSHs (*av_fetch (av, 0, 1));
871	PUSHs (sv_cbor);	1007	PUSHs (sv_cbor);
872		1008
…		…
896		1032
897	default:	1033	default:
898	{	1034	{
899	sv = decode_sv (dec);	1035	sv = decode_sv (dec);
900		1036
		1037	dSP;
		1038	ENTER; SAVETMPS;
		1039	SAVESTACK_POS ();
		1040	PUSHMARK (SP);
		1041	EXTEND (SP, 2);
		1042	PUSHs (newSVuv (tag));
		1043	PUSHs (sv);
		1044
		1045	PUTBACK;
		1046	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);
		1047	SPAGAIN;
		1048
		1049	if (SvTRUE (ERRSV))
		1050	{
		1051	FREETMPS; LEAVE;
		1052	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
		1053	}
		1054
		1055	if (count)
		1056	{
		1057	SvREFCNT_dec (sv);
		1058	sv = SvREFCNT_inc (POPs);
		1059	}
		1060	else
		1061	{
901	AV *av = newAV ();	1062	AV *av = newAV ();
902	av_push (av, newSVuv (tag));	1063	av_push (av, newSVuv (tag));
903	av_push (av, sv);	1064	av_push (av, sv);
904		1065
905	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1066	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
906	? cbor_tagged_stash	1067	? cbor_tagged_stash
907	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1068	: gv_stashpv ("CBOR::XS::Tagged" , 1);
908
909	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);	1069	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
		1070	}
		1071
		1072	PUTBACK;
		1073
		1074	FREETMPS; LEAVE;
910	}	1075	}
911	break;	1076	break;
912	}	1077	}
913		1078
914	return sv;	1079	return sv;
…		…
921	static SV *	1086	static SV *
922	decode_sv (dec_t *dec)	1087	decode_sv (dec_t *dec)
923	{	1088	{
924	WANT (1);	1089	WANT (1);
925		1090
926	switch (*dec->cur >> 5)	1091	switch (*dec->cur >> MAJOR_SHIFT)
927	{	1092	{
928	case 0: // unsigned int	1093	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
929	return newSVuv (decode_uint (dec));	1094	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
930	case 1: // negative int	1095	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
931	return newSViv (-1 - (IV)decode_uint (dec));	1096	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
932	case 2: // octet string	1097	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
933	return decode_str (dec, 0);	1098	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
934	case 3: // utf-8 string	1099	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
935	return decode_str (dec, 1);	1100
936	case 4: // array	1101	case MAJOR_MISC >> MAJOR_SHIFT:
937	return decode_av (dec);
938	case 5: // map
939	return decode_hv (dec);
940	case 6: // tag
941	return decode_tagged (dec);
942	case 7: // misc
943	switch (*dec->cur++ & 31)	1102	switch (*dec->cur++ & MINOR_MASK)
944	{	1103	{
945	case 20:	1104	case SIMPLE_FALSE:
946	#if CBOR_SLOW	1105	#if CBOR_SLOW
947	types_false = get_bool ("Types::Serialiser::false");	1106	types_false = get_bool ("Types::Serialiser::false");
948	#endif	1107	#endif
949	return newSVsv (types_false);	1108	return newSVsv (types_false);
950	case 21:	1109	case SIMPLE_TRUE:
951	#if CBOR_SLOW	1110	#if CBOR_SLOW
952	types_true = get_bool ("Types::Serialiser::true");	1111	types_true = get_bool ("Types::Serialiser::true");
953	#endif	1112	#endif
954	return newSVsv (types_true);	1113	return newSVsv (types_true);
955	case 22:	1114	case SIMPLE_NULL:
956	return newSVsv (&PL_sv_undef);	1115	return newSVsv (&PL_sv_undef);
957	case 23:	1116	case SIMPLE_UNDEF:
958	#if CBOR_SLOW	1117	#if CBOR_SLOW
959	types_error = get_bool ("Types::Serialiser::error");	1118	types_error = get_bool ("Types::Serialiser::error");
960	#endif	1119	#endif
961	return newSVsv (types_error);	1120	return newSVsv (types_error);
962		1121
963	case 25:	1122	case MISC_FLOAT16:
964	{	1123	{
965	WANT (2);	1124	WANT (2);
966		1125
967	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1126	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
968	dec->cur += 2;	1127	dec->cur += 2;
969		1128
970	return newSVnv (ecb_binary16_to_float (fp));	1129	return newSVnv (ecb_binary16_to_float (fp));
971	}	1130	}
972		1131
973	case 26:	1132	case MISC_FLOAT32:
974	{	1133	{
975	uint32_t fp;	1134	uint32_t fp;
976	WANT (4);	1135	WANT (4);
977	memcpy (&fp, dec->cur, 4);	1136	memcpy (&fp, dec->cur, 4);
978	dec->cur += 4;	1137	dec->cur += 4;
…		…
981	fp = ecb_bswap32 (fp);	1140	fp = ecb_bswap32 (fp);
982		1141
983	return newSVnv (ecb_binary32_to_float (fp));	1142	return newSVnv (ecb_binary32_to_float (fp));
984	}	1143	}
985		1144
986	case 27:	1145	case MISC_FLOAT64:
987	{	1146	{
988	uint64_t fp;	1147	uint64_t fp;
989	WANT (8);	1148	WANT (8);
990	memcpy (&fp, dec->cur, 8);	1149	memcpy (&fp, dec->cur, 8);
991	dec->cur += 8;	1150	dec->cur += 8;
…		…
994	fp = ecb_bswap64 (fp);	1153	fp = ecb_bswap64 (fp);
995		1154
996	return newSVnv (ecb_binary64_to_double (fp));	1155	return newSVnv (ecb_binary64_to_double (fp));
997	}	1156	}
998		1157
999	// 0..19 unassigned	1158	// 0..19 unassigned simple
1000	// 24 reserved + unassigned (reserved values are not encodable)	1159	// 24 reserved + unassigned simple (reserved values are not encodable)
		1160	// 28-30 unassigned misc
		1161	// 31 break code
1001	default:	1162	default:
1002	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1163	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1003	}	1164	}
1004		1165
1005	break;	1166	break;
1006	}	1167	}
1007		1168
…		…
1010	}	1171	}
1011		1172
1012	static SV *	1173	static SV *
1013	decode_cbor (SV string, CBOR cbor, char **offset_return)	1174	decode_cbor (SV string, CBOR cbor, char **offset_return)
1014	{	1175	{
1015	dec_t dec = { };	1176	dec_t dec = { 0 };
1016	SV *sv;	1177	SV *sv;
1017	STRLEN len;	1178	STRLEN len;
1018	char *data = SvPVbyte (string, len);	1179	char *data = SvPVbyte (string, len);
1019		1180
1020	if (len > cbor->max_size && cbor->max_size)	1181	if (len > cbor->max_size && cbor->max_size)
…		…
1034	if (dec.cur != dec.end && !dec.err)	1195	if (dec.cur != dec.end && !dec.err)
1035	dec.err = "garbage after CBOR object";	1196	dec.err = "garbage after CBOR object";
1036		1197
1037	if (dec.err)	1198	if (dec.err)
1038	{	1199	{
		1200	if (dec.shareable)
		1201	{
		1202	// need to break cyclic links, which whould all be in shareable
		1203	int i;
		1204	SV **svp;
		1205
		1206	for (i = av_len (dec.shareable) + 1; i--; )
		1207	if ((svp = av_fetch (dec.shareable, i, 0)))
		1208	sv_setsv (*svp, &PL_sv_undef);
		1209	}
		1210
1039	SvREFCNT_dec (sv);	1211	SvREFCNT_dec (sv);
1040	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1212	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1041	}	1213	}
1042		1214
1043	sv = sv_2mortal (sv);	1215	sv = sv_2mortal (sv);
1044		1216
1045	return sv;	1217	return sv;
1046	}	1218	}
1047		1219
		1220	/////////////////////////////////////////////////////////////////////////////
		1221	// incremental parser
		1222
		1223	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1224
		1225	// returns 0 for notyet, 1 for success or error
		1226	static int
		1227	incr_parse (CBOR self, SV cborstr)
		1228	{
		1229	STRLEN cur;
		1230	SvPV (cborstr, cur);
		1231
		1232	while (ecb_expect_true (self->incr_need <= cur))
		1233	{
		1234	// table of integer count bytes
		1235	static I8 incr_len[MINOR_MASK + 1] = {
		1236	0, 0, 0, 0, 0, 0, 0, 0,
		1237	0, 0, 0, 0, 0, 0, 0, 0,
		1238	0, 0, 0, 0, 0, 0, 0, 0,
		1239	1, 2, 4, 8,-1,-1,-1,-2
		1240	};
		1241
		1242	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1243	U8 m = *p & MINOR_MASK;
		1244	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1245	I8 ilen = incr_len[m];
		1246
		1247	self->incr_need = self->incr_pos + 1;
		1248
		1249	if (ecb_expect_false (ilen < 0))
		1250	{
		1251	if (m != MINOR_INDEF)
		1252	return 1; // error
		1253
		1254	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1255	{
		1256	if (count >= 0)
		1257	return 1; // error
		1258
		1259	count = 1;
		1260	}
		1261	else
		1262	{
		1263	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1264	count = -1;
		1265	}
		1266	}
		1267	else
		1268	{
		1269	self->incr_need += ilen;
		1270	if (ecb_expect_false (self->incr_need > cur))
		1271	return 0;
		1272
		1273	int major = *p >> MAJOR_SHIFT;
		1274
		1275	switch (major)
		1276	{
		1277	case MAJOR_TAG >> MAJOR_SHIFT:
		1278	++count; // tags merely prefix another value
		1279	break;
		1280
		1281	case MAJOR_BYTES >> MAJOR_SHIFT:
		1282	case MAJOR_TEXT >> MAJOR_SHIFT:
		1283	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1284	case MAJOR_MAP >> MAJOR_SHIFT:
		1285	{
		1286	UV len;
		1287
		1288	if (ecb_expect_false (ilen))
		1289	{
		1290	len = 0;
		1291
		1292	do {
		1293	len = (len << 8) \| *++p;
		1294	} while (--ilen);
		1295	}
		1296	else
		1297	len = m;
		1298
		1299	switch (major)
		1300	{
		1301	case MAJOR_BYTES >> MAJOR_SHIFT:
		1302	case MAJOR_TEXT >> MAJOR_SHIFT:
		1303	self->incr_need += len;
		1304	if (ecb_expect_false (self->incr_need > cur))
		1305	return 0;
		1306
		1307	break;
		1308
		1309	case MAJOR_MAP >> MAJOR_SHIFT:
		1310	len <<= 1;
		1311	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1312	if (len)
		1313	{
		1314	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1315	count = len + 1;
		1316	}
		1317	break;
		1318	}
		1319	}
		1320	}
		1321	}
		1322
		1323	self->incr_pos = self->incr_need;
		1324
		1325	if (count > 0)
		1326	{
		1327	while (!--count)
		1328	{
		1329	if (!AvFILLp (self->incr_count))
		1330	return 1; // done
		1331
		1332	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1333	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1334	}
		1335
		1336	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1337	}
		1338	}
		1339
		1340	return 0;
		1341	}
		1342
		1343
1048	/////////////////////////////////////////////////////////////////////////////	1344	/////////////////////////////////////////////////////////////////////////////
1049	// XS interface functions	1345	// XS interface functions
1050		1346
1051	MODULE = CBOR::XS PACKAGE = CBOR::XS	1347	MODULE = CBOR::XS PACKAGE = CBOR::XS
1052		1348
…		…
1059	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);	1355	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1060		1356
1061	types_true = get_bool ("Types::Serialiser::true" );	1357	types_true = get_bool ("Types::Serialiser::true" );
1062	types_false = get_bool ("Types::Serialiser::false");	1358	types_false = get_bool ("Types::Serialiser::false");
1063	types_error = get_bool ("Types::Serialiser::error");	1359	types_error = get_bool ("Types::Serialiser::error");
		1360
		1361	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1064		1362
1065	sv_cbor = newSVpv ("CBOR", 0);	1363	sv_cbor = newSVpv ("CBOR", 0);
1066	SvREADONLY_on (sv_cbor);	1364	SvREADONLY_on (sv_cbor);
1067	}	1365	}
1068		1366
…		…
1090	void shrink (CBOR *self, int enable = 1)	1388	void shrink (CBOR *self, int enable = 1)
1091	ALIAS:	1389	ALIAS:
1092	shrink = F_SHRINK	1390	shrink = F_SHRINK
1093	allow_unknown = F_ALLOW_UNKNOWN	1391	allow_unknown = F_ALLOW_UNKNOWN
1094	allow_sharing = F_ALLOW_SHARING	1392	allow_sharing = F_ALLOW_SHARING
1095	dedup_keys = F_DEDUP_KEYS	1393	allow_cycles = F_ALLOW_CYCLES
1096	dedup_strings = F_DEDUP_STRINGS	1394	pack_strings = F_PACK_STRINGS
		1395	validate_utf8 = F_VALIDATE_UTF8
1097	PPCODE:	1396	PPCODE:
1098	{	1397	{
1099	if (enable)	1398	if (enable)
1100	self->flags \|= ix;	1399	self->flags \|= ix;
1101	else	1400	else
…		…
1107	void get_shrink (CBOR *self)	1406	void get_shrink (CBOR *self)
1108	ALIAS:	1407	ALIAS:
1109	get_shrink = F_SHRINK	1408	get_shrink = F_SHRINK
1110	get_allow_unknown = F_ALLOW_UNKNOWN	1409	get_allow_unknown = F_ALLOW_UNKNOWN
1111	get_allow_sharing = F_ALLOW_SHARING	1410	get_allow_sharing = F_ALLOW_SHARING
1112	get_dedup_keys = F_DEDUP_KEYS	1411	get_allow_cycles = F_ALLOW_CYCLES
1113	get_dedup_strings = F_DEDUP_STRINGS	1412	get_pack_strings = F_PACK_STRINGS
		1413	get_validate_utf8 = F_VALIDATE_UTF8
1114	PPCODE:	1414	PPCODE:
1115	XPUSHs (boolSV (self->flags & ix));	1415	XPUSHs (boolSV (self->flags & ix));
1116		1416
1117	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1417	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1118	PPCODE:	1418	PPCODE:
…		…
1131	XPUSHs (ST (0));	1431	XPUSHs (ST (0));
1132		1432
1133	int get_max_size (CBOR *self)	1433	int get_max_size (CBOR *self)
1134	CODE:	1434	CODE:
1135	RETVAL = self->max_size;	1435	RETVAL = self->max_size;
		1436	OUTPUT:
		1437	RETVAL
		1438
		1439	void filter (CBOR self, SV filter = 0)
		1440	PPCODE:
		1441	SvREFCNT_dec (self->filter);
		1442	self->filter = filter ? newSVsv (filter) : filter;
		1443	XPUSHs (ST (0));
		1444
		1445	SV get_filter (CBOR self)
		1446	CODE:
		1447	RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1136	OUTPUT:	1448	OUTPUT:
1137	RETVAL	1449	RETVAL
1138		1450
1139	void encode (CBOR self, SV scalar)	1451	void encode (CBOR self, SV scalar)
1140	PPCODE:	1452	PPCODE:
…		…
1155	EXTEND (SP, 2);	1467	EXTEND (SP, 2);
1156	PUSHs (sv);	1468	PUSHs (sv);
1157	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1469	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1158	}	1470	}
1159		1471
		1472	void incr_parse (CBOR self, SV cborstr)
		1473	ALIAS:
		1474	incr_parse_multiple = 1
		1475	PPCODE:
		1476	{
		1477	if (SvUTF8 (cborstr))
		1478	sv_utf8_downgrade (cborstr, 0);
		1479
		1480	if (!self->incr_count)
		1481	{
		1482	self->incr_count = newAV ();
		1483	self->incr_pos = 0;
		1484	self->incr_need = 1;
		1485
		1486	av_push (self->incr_count, newSViv (1));
		1487	}
		1488
		1489	do
		1490	{
		1491	if (!incr_parse (self, cborstr))
		1492	{
		1493	if (self->incr_need > self->max_size && self->max_size)
		1494	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1495	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1496
		1497	break;
		1498	}
		1499
		1500	SV *sv;
		1501	char *offset;
		1502
		1503	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1504	XPUSHs (sv);
		1505
		1506	sv_chop (cborstr, offset);
		1507
		1508	av_clear (self->incr_count);
		1509	av_push (self->incr_count, newSViv (1));
		1510
		1511	self->incr_pos = 0;
		1512	self->incr_need = self->incr_pos + 1;
		1513	}
		1514	while (ix);
		1515	}
		1516
		1517	void incr_reset (CBOR *self)
		1518	CODE:
		1519	{
		1520	SvREFCNT_dec (self->incr_count);
		1521	self->incr_count = 0;
		1522	}
		1523
		1524	void DESTROY (CBOR *self)
		1525	PPCODE:
		1526	cbor_free (self);
		1527
1160	PROTOTYPES: ENABLE	1528	PROTOTYPES: ENABLE
1161		1529
1162	void encode_cbor (SV *scalar)	1530	void encode_cbor (SV *scalar)
		1531	ALIAS:
		1532	encode_cbor = 0
		1533	encode_cbor_sharing = F_ALLOW_SHARING
1163	PPCODE:	1534	PPCODE:
1164	{	1535	{
1165	CBOR cbor;	1536	CBOR cbor;
1166	cbor_init (&cbor);	1537	cbor_init (&cbor);
		1538	cbor.flags \|= ix;
1167	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1539	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1168	XPUSHs (scalar);	1540	XPUSHs (scalar);
1169	}	1541	}
1170		1542
1171	void decode_cbor (SV *cborstr)	1543	void decode_cbor (SV *cborstr)

Diff Legend

-–
+Removed lines
-+
+Added lines
-<
+Changed lines
->
+Changed lines

Comparing CBOR-XS/XS.xs (file contents): Revision 1.20 by root, Wed Nov 20 11:06:42 2013 UTC vs. Revision 1.50 by root, Thu Feb 25 02:29:22 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.20 by root, Wed Nov 20 11:06:42 2013 UTC vs.
Revision 1.50 by root, Thu Feb 25 02:29:22 2016 UTC